System for allocation of network resources using an autoregressive integrated moving average method

A method for allocating network resources to a first computer system requesting a communications path to a second computer system, the method comprising the steps of:

• receiving a request from the first computer system for the communications path to the second computer system;
• computing a prediction if the network resources will have enough bandwidth capacity for establishing the communications path;
• and
• granting or denying the request as a function of the prediction, wherein the network resources carry conforming and non-conforming traffic, and wherein the communications path is to be part of the non-conforming traffic within the network resources, and wherein the bandwidth capacity is a predicted maximum amount of bandwidth available in the network resources for the non-conforming traffic, and wherein the communications path has an amount of bandwidth, and wherein the request will be granted by the connection agent when the bandwidth capacity is greater than or equal to the amount of bandwidth.

A method for allocating network resources to a first computer system requesting a communications path to a second computer system, the method comprising the steps of;

• receiving a request from the first computer system for the communications path to the second computer system:
• computing a prediction if the network resources will have enough bandwidth capacity for establishing the communications path;
• granting or denying the request as a function of the prediction;
• if the computing step predicts that there will not be enough bandwidth capacity for establishing the communications path, computing a prediction if the network resources will have enough bandwidth capacity for establishing the communications path at a future time t;
• and
• granting the request if the prediction of network resources at the future time t indicates that there will be enough bandwidth capacity for establishing the communications path at the future time t.

A data processing network comprising a plurality of client computer systems coupled together to one or more application server systems, wherein the data processing network further comprises a connection control system for allocating network resources to a first client computer system requesting a communications path to an applications server, the connection control system comprising:

• a connection agent for receiving a request from the first client computer system for the communications path to the application server;
• a bandwidth predictor operable or computing a prediction if the network resources will have enough bandwidth capacity for establishing the communications path, wherein the connection agent grants or denies the request as a function of the prediction, wherein the bandwidth predictor uses an autoregressive integrated moving averages trend analyses to compute the prediction;
• and
• a static model generation utility coupled to the bandwidth predictor for analyzing bandwidth utilization over one or more periods of time and for generating coefficients as a result of the bandwidth utilization analyzing to be input into the autoregressive integrated moving averages trend analyses,
• wherein the communications path has an amount of bandwidth, and wherein the request will be granted by the connection agent when the bandwidth capacity is greater than or equal to the amount of bandwidth.